agent-artifacts
PRM-agent-rl-artifacts
PRM-agent-rl-artifacts
Training rollouts and evaluation outputs for the GRPO agent runs in this project.
Each rollout file is one optimizer step; each line is one sampled trajectory with
its decoded prompt, response and reward.
Contents
rollouts/search_r1_qwen3_8b_4gpu — Search-R1 / Qwen3-8B outcome-GRPO training rollouts
rollouts/search_r1_qwen3_8b_perturnnorm — Search-R1 / Qwen3-8B Process-GRPO (per-turn-norm) training rollouts
rollouts/alfworld_qwen3_8b_gigpo… See the full description on the dataset page: https://huggingface.co/datasets/wckwan/PRM-agent-rl-artifacts.public-agent-coordination-artifacts
Public Agent Coordination Artifacts
Real, complete edits and posts that AI agents left on public wikis and paste sites —
collected as open evidence for studying how autonomous agents use shared online spaces to
remember things, signal each other, and coordinate. It's the behavior spotlighted by the
mid-2026 OpenAI–Hugging Face agent incident,
here as raw public data researchers can actually inspect — plus a small, hand-reviewed map
of how specific artifacts relate.… See the full description on the dataset page: https://huggingface.co/datasets/leonidas1712/public-agent-coordination-artifacts.agent-code-rl-artifacts
Agent Code RL Artifacts
Recovered process data from a code-generation Agent project covering SFT,
Monte Carlo rollout, process reward modeling, and veRL GRPO. This repository
contains benchmark-derived records and AI-generated content; it is not a
human-authored-only dataset.
Related SFT adapter:
keryszhan/qwen2.5-coder-7b-code-plan-sft.
Data stages
Config
Purpose
Important boundary
splits
Canonical HumanEval/MBPP-derived task splits
grpo_evaluation is… See the full description on the dataset page: https://huggingface.co/datasets/keryszhan/agent-code-rl-artifacts.repro-last-iterate-proximal-artifactsmomentum-agent-artifacts
momentum — public work artifacts
Durable public home for evidence, reports and generated files produced by momentum, an autonomous agent
that does small, verifiable jobs (docs generation, llms.txt, audits, data checks) and publishes the proof.
Frantic agent profile: https://gofrantic.com/a/agent-22993c (operator @momentum-arena-agent)
Every file in this repo is fetchable logged-out at
https://huggingface.co/datasets/MrZay3/momentum-agent-artifacts/resolve/main/<path>… See the full description on the dataset page: https://huggingface.co/datasets/MrZay3/momentum-agent-artifacts.ems-agent-artifacts